How to make a filter more efficient

Question

I've created a script to compare a list of users from a MSSQL database against a list of disabled AD users. The tool we use for AD authentication returns data in the following format (with plenty more whitespace in between the columns), so to allow the compare (in python) the data needs to be parsed:

User info (Level-2):
====================
Name:                         domain\username
UPN:                          [email protected]
Generated UPN:                NO
DN:                           DN....
Uid:                          123456
Gid:                          123456
Gecos:                        User Name
Shell:                        /shell/path
Home dir:                     /homedir
Other attributes:         ....
Other attributes:         ....
Account disabled (or locked): TRUE
Other attributes:         ....
Other attributes:         ....
Other attributes:         ....
etc

From the command line, the following does the trick, however it's very messy and I'm hoping someone may be able to suggest a cleaner method. I've experimented with awk but haven't had much success in reducing the size of the command:

/opt/util/enum-users --level 2 |
grep "Name:\|Account disabled (or locked): TRUE" |
grep -x "Account disabled (or locked): TRUE" -B 1 |
grep "Name:" |
sed -r 's/^.{35}//'

This outputs the data as required:

username
username
etc

muru · Accepted Answer · 2015-01-06 16:36:22Z

1

I don't know if it is more efficient, but you could do this entirely in awk:

awk -F': *' '$1 == "Name" {name = $2; next}; 
$0 == "Account disabled (or locked): TRUE" {
   print gensub(/.*\\/, "", 1, name)
}'

edited Jan 6, 2015 at 16:36

answered Jan 5, 2015 at 12:07

muru

78.4k16 gold badges214 silver badges320 bronze badges

clever use of awk separators. +1

Olivier Dulac
– Olivier Dulac

2015-01-05 13:59:53 +00:00
Commented Jan 5, 2015 at 13:59

Add a comment |

mikeserv · Accepted Answer · 2015-01-05 23:50:57Z

0

You should just do this with sed entirely since you're using it anyway:

/opt/util/enum-users --level 2 |
sed -n '/^Name:/h;//,/^Account/{
    /^Account/!d;/TRUE[^:]*$/!d;g
    s/^[^:]*:[[:blank:]]*//p
}'

I don't actually know how the [[:blank:]] business should be handled, but the assumption made above is that an interesting line group will probably start like:

^Name:[[:blank:]]*[INFORMATION YOU WANT]$

...and that the next occurring line beginning with the string Account should end with the string TRUE followed by any number of blanks and that an uninteresting group will probably have FALSE or something.

answered Jan 5, 2015 at 23:50

mikeserv

59.4k10 gold badges123 silver badges244 bronze badges

Thanks mikeserv. The only thing this didn't do was strip the "domain\", however it was a great help.

Paul Dundee
– Paul Dundee

2015-01-06 13:35:37 +00:00
Commented Jan 6, 2015 at 13:35
@PaulDundee - I wasn't aware it should strip the domain. If that is the case you can just add a s/@.*// after the substitution statement already there. Oh, but move the p.

mikeserv
– mikeserv

2015-01-06 13:46:00 +00:00
Commented Jan 6, 2015 at 13:46

Add a comment |

Stack Exchange Network

How to make a filter more efficient

2 Answers 2

You must log in to answer this question.

Hot Network Questions

How to make a filter more efficient

2 Answers 2

You must log in to answer this question.

Related

Hot Network Questions